Does social deprivation index (SDI) portend predictive value of perioperative outcomes after adult spinal deformity (ASD) surgery?

BACKGROUND CONTEXT: The ASD population is extremely heterogeneous, and the condition is not confined to any one sociodemographic class. Still, considerable disparities exist in equitable distribution of health care resources and availability of interventions and postoperative supportive resources. PURPOSE: We explore the relationship of Social Deprivation Index (SDI) to initial presentation and perioperative outcomes after ASD surgery to better understand the compounding effects of social disparities on patients’ trajectories. STUDY DESIGN/SETTING: Retrospective cohort study of multi-center prospectively enrolled US database. PATIENT SAMPLE: A total of 874 ASD patients. OUTCOME MEASURES: Complications, alignment, HRQLS, discharge dispositions. METHODS: Operative patients with complete 6-week data were queried. SDI scores were derived from Zip Code Tabulation Areas data. The primary outcome was severity at initial presentation. Patients were ranked by quartile, and Q1, Q2, and Q3 were grouped as nondeprived (ND) while the deprived (D) cohort was composed of patients in Q4 (SDI>75^th percentile). Means comparison tests, univariate analysis. pairwise correlation matrices were applied to identify trends and associations between SDI and perioperative outcomes. RESULTS: A total of 579 patients met inclusion criteria (mean age: 60.6±15.6 yrs, 69.4% female, mean BMI: 27.3±5.5 kg/m2, CCI: 0.95±1.5, operative time: 422±149.3 minutes, EBL: 1438.5±1207.7 mL, LOS: 7.3±5.4 days. The mean SDI was 37.3±27.7 out of 100, with a mean poverty subscore of 37.8±28.6. 142 (25.6%) patients were Q1, 135 (24.4%) were Q2, 140 (25.3%) were Q3, and 137 (24.7%) were Q4. There was a significant difference between poverty-level sub-scores (ND: 35.6 vs D: 74.9, p<.001). While there were no significant differences amongst the baseline demographics, there were significant differences between key medical history components; D patients demonstrated greater rates of a history of myocardial infarction (6.6% vs 2.9%, p=0.05) and lung disease (9.5% vs 3.4%, p=0.004), while ND patients carried higher rates of osteoporosis than D (23.7% v. 14.6%, p=0.024). There was also a significant difference in BL frailty (ND: 3.15 vs D: 3.7, p =0.027). D patients were significantly more likely to be considered frail at baseline per Edmonton frailty scores (OR 2.9, CI 95% 1.5-5.9, p=0.003). D patients were also less likely to be optimized in terms of receiving medications for osteoporosis (OR 0.4, CI 95% 0.24-0.80m p=0.007). In terms of BL HRQLs, ND patients had better VR-12 Mental scores (p=0.17) ODI–Walking subscores (p=0.02). In a sub-analysis of those who were considered extremely deprived (>90th percentile), BL HRQLs demonstrated even greater pain and disability in terms of multiple ODI domains and worse baseline health status (p<0.05, all). Preoperatively, there were no significant differences in radiographic parameters, GAP categories, or Schwab modifiers. There were also significant differences in discharge dispositions, with the deprived (D) group being discharged to home at a higher rate and entering rehab at a lower rate (p=0.027). CONCLUSIONS: Patients considered “socially deprived” may experience worse perioperative outcomes in terms of complications and alignment. Resources should be allocated towards building more substantial support systems of such patients in this vulnerable period. FDA Device/Drug Status: This abstract does not discuss or include any applicable devices or drugs.